1- Cocaine seizures

library(ggplot2)
data("cocaine", package = "ggvis")
str(cocaine)
summary(cocaine)

• state: the state in which the seizure occured • potency: the purity of cocaine (as percentage) • weight: the weight (in grams) of the seized cocaine • month: the months in which the seizure occured • price: the estimated value (in USD).

cocaine2 = subset(cocaine, weight < 200)
ggplot(cocaine2, aes(x = state)) + geom_bar() + theme(axis.text.x = element_text(angle = 90, hjust = 1))

# Florida (FL), New York (NY) and Virginia (VA) have the highest number of seizures
reorder_size <- function(x) {
  factor(x, levels = names(sort(table(x), decreasing=TRUE)))
}
ggplot(cocaine2, aes(reorder_size(state))) + geom_bar() + theme(axis.text.x = element_text(angle = 90, hjust = 1))

ggplot(cocaine2, aes(x = weight)) + geom_histogram()
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

# It is more likely that smaller weight happen more 
ggplot(cocaine2, aes(x = potency, y = price)) + geom_point(alpha = 0.3) 

# I do not see a clear relationshio netween price and potency
ggplot(cocaine2, aes(x = weight, y = price)) + geom_point(alpha = 0.3)

# It seems that there is a linear relationshio netween price and weight
ggplot(cocaine2, aes(x = potency, y = price, color = weight)) + geom_point(alpha = 0.5) + scale_color_gradient(trans = "log") 

ggplot(cocaine2, aes(x = weight, y = potency)) + geom_point(alpha = 0.3) + geom_smooth()

# It seems that the average potency of lower weight cocaine is higher (considering weights below 150 grams)

2- Flights data

require(nycflights13)
#hour, minute: the hour and minute of the departure
#arr delay: the arrival delay of the incoming plane (in minutes)
#dest: the destination.
flights$time <- flights$hour + flights$minute / 60
delay.per.hour <- aggregate(formula = arr_delay ~ time, data = flights, FUN = mean, na.rm = TRUE)
ggplot(data = delay.per.hour, aes(time, arr_delay)) + geom_point()

# It seems that arrival delay gradually increases from mornig 5am to 12am, and the sudenly we have several hour delays after midnight. So, it's better to take a flight in the morning.
delay.per.hour$n = aggregate(formula = arr_delay ~ time, data = flights, FUN = length)[, "arr_delay"]
ggplot(data = delay.per.hour, aes(time, arr_delay, size = n)) + geom_point(alpha = 0.3)

# this plot is more informative because we can see the nmber of plots is different hours.For example, the number of fights after midnight is mush less than the other hours that cannot be understood in the previous plot.
delay.per.hour.subset = subset(delay.per.hour, n > 50)
ggplot(data = delay.per.hour.subset, aes(time, arr_delay, size = n)) + geom_point() + scale_size_area()

library(dplyr)
delay.per.hour = summarise(group_by(flights, time), arr_delay = mean(arr_delay, na.rm = TRUE))

3- Flights data, continued

require(nycflights13)
delay.per.dest = aggregate(formula = arr_delay ~ dest, data = flights,FUN = mean, na.rm = TRUE)
delay.per.dest$n =  aggregate(formula = arr_delay ~ dest, data = flights, FUN = length)[, "arr_delay"]
delay.per.dest = merge(x = delay.per.dest, y = airports, by.x = "dest", by.y = "faa", all.x = T, all.y = F)
ggplot(data = delay.per.dest, aes(lon, lat, size = n)) + geom_point(alpha = 0.5)

ggplot(data = delay.per.dest, aes(lon, lat, size = n, color = arr_delay)) + geom_point()

#It seems that there is not any relationship between average arrival delay and number of departing planes
ggplot(data = delay.per.dest, aes(lon, lat, size = n, color = arr_delay)) + geom_point() + scale_size_area()

require(maps)
delay.per.dest.subset = subset(delay.per.dest, lon > -140)
ggplot(data = delay.per.dest.subset, aes(lon, lat, size = n, color = arr_delay)) + geom_point() + borders(database = "state", size = 0.5) + scale_size_area()

4- Gapminder: Fact-based world view

require(gapminder, quietly = TRUE)

ggplot(subset(gapminder, year == 2002), aes(x = gdpPercap, y = lifeExp)) + geom_point()

ggplot(subset(gapminder, year == 2002), aes(x = gdpPercap, y = lifeExp, fill = country, )) + geom_point(aes(size = sqrt(pop)), pch = 21, show_guide = FALSE)

ggplot(subset(gapminder, year == 2002), aes(x = gdpPercap, y = lifeExp, fill = country, )) + geom_point(aes(size = sqrt(pop)), pch = 21, show_guide = FALSE) + scale_x_log10()

ggplot(subset(gapminder, year == 2002), aes(x = gdpPercap, y = lifeExp, fill = country, )) + geom_point(aes(size = sqrt(pop)), pch = 21, show_guide = FALSE) + scale_x_log10() + scale_size_continuous(range = c(1,40))

ggplot(subset(gapminder, year == 2002), aes(x = gdpPercap, y = lifeExp, fill = country, )) + geom_point(aes(size = sqrt(pop)), pch = 21, show_guide = FALSE) + scale_x_log10() + scale_size_continuous(range = c(1,40)) + scale_fill_manual(values = country_colors)

ggplot(subset(gapminder, year == 2002), aes(x = gdpPercap, y = lifeExp, fill = country, )) + geom_point(aes(size = sqrt(pop)), pch = 21, show_guide = FALSE) + scale_x_log10() + scale_size_continuous(range = c(1,40)) + scale_fill_manual(values = country_colors) + facet_grid( . ~ continent)

5- Napoleon

data("Minard.troops", package = "HistData")
str(Minard.troops)
## 'data.frame':    51 obs. of  5 variables:
##  $ long     : num  24 24.5 25.5 26 27 28 28.5 29 30 30.3 ...
##  $ lat      : num  54.9 55 54.5 54.7 54.8 54.9 55 55.1 55.2 55.3 ...
##  $ survivors: int  340000 340000 340000 320000 300000 280000 240000 210000 180000 175000 ...
##  $ direction: Factor w/ 2 levels "A","R": 1 1 1 1 1 1 1 1 1 1 ...
##  $ group    : int  1 1 1 1 1 1 1 1 1 1 ...
data("Minard.cities", package = "HistData")
str(Minard.cities)
## 'data.frame':    20 obs. of  3 variables:
##  $ long: num  24 25.3 26.4 26.8 27.7 27.6 28.5 28.7 29.2 30.2 ...
##  $ lat : num  55 54.7 54.4 54.3 55.2 53.9 54.3 55.5 54.4 55.3 ...
##  $ city: Factor w/ 20 levels "Bobr","Chjat",..: 5 18 15 9 4 7 16 13 1 19 ...
plot_troops = ggplot(Minard.troops, aes(x = long, y = lat)) + geom_path(alpha = 0.6, aes(size = survivors, colour = direction, group = group), lineend = "round")
plot_troops

plot_both = plot_troops + geom_text(aes(label = city), size = 3, data = Minard.cities)
plot_both

plot_polished = plot_both + scale_color_manual(values = c("red", "grey")) + xlab(NULL) + ylab(NULL)
plot_polished

library(knitr) rmarkdown::render(‘ghandehari_lab06.Rmd’)